Oddjob

Van,

You thinking its an electronic fault, or that the car is actually overboosting?

A bad knock sensor can cause a KLR diagnostic fault/blink code and make the car run in limp mode - likely shouldnt cause the car to hit the DME overboost fuel cutoff. Im fairly certain that function is independent of the KLR (and the knock sensor).

You using the stock/factory DME chip to stay in class? Have you tried an aftermarket DME chip to see if the problem goes away? If the problem is gone, then it likely indicates that the car is actually running too much boost (for the factory chip settings) or that the AFM is sending a faulty/inaccurate signal to the DME. If the problem is still there with the aftermarket chip, then I would continue with the diagnostic checks on the sensors and ECUs.

Van

I replaced the cycling valve a few years ago, because it had gone bad. A new one fixed it. When this happened again, the cycling valve was the first thing that came to mind - though I was surprised that it had only lasted a few years.

I put a new one it, but it didn't solve the problem (now I have a spare, good cycling valve!)


Perhaps I don't know what the difference between limp mode and overboost fuel cut off are... I've been referring to them as the same thing.

In the service manual (28-30) when it's talking about the KLR's blink codes, it says: "The knock/charging pressure system includes a self-monitoring system. If a defect has been found, the engine will operate on a safety system (basic charging pressure up to approx. 0.3 bar overpressure and 6 degree later ignition), which means significantly reduced performance."

Testing the knock sensor (KLR plug pins 11 and 13 - and then directly on the sensor itself), shows it is clearly bad. I have a new one on order ($57) and will report my findings once I have it installed.

It makes logical sense to me that the KLR only looks for knock at WOT and full boost (when else would knock occur if your car is running right?) When the KLR realizes it's not getting a signal from the knock sensor, it goes into this safety mode, primarily because it can retard the timing.

Oddjob

Overboost protection is a DME fuel cutoff. Airflow is measured by the AFM and the DME determines if it exceeds a preset limit through an rpm range, mapped in the DME chip. Aftermarket DME chips raise/eliminate this factory protection, so you can increase boost w/o hitting the fuel cutoff. When you hit this, its a very dramatic stumble, almost violent. Car feels like it just fell on its face. Usually hits in the upper rpms 4500 to 5k or so, in 3rd gear and above. After you hit this fuel cutoff, any throttle position greater than about 3/4's will trip the fuel cutoff again within about 60 seconds (it automatically resets after so many firing cycles if the overboost condition is not detected again) - but you can also reset the DME by shutting the ignition off, which is often "necessary" to do on the fly when on track/racing.

The limp home mode is a KLR fault detection/protection. Various faults will cause the KLR to reduce boost (through the CV) and tell the DME to retard timing. The CV basically runs open, keeping the WG open and the car will only generate approx 1.2-1.3 bar boost. Car runs normal in all other aspects, just short on power, feels like driving an N/A at full throttle. If in doubt, unplug the TPS and drive around. Other than the high/hunting idle, this is what the limp mode will feel like if the KLR has found other faults with sensors. Also can get the blink code faults from the diagnostic port behind the driver side shock tower. Something like 2-1, 2-2 blink codes are the knock sensor faults and 3-1, 3-2 are the low or high charge pressure faults (note: the KLR excessive boost pressure diagnostic fault is not the same as the DME overboost protection). The KLR runs dianostic checks on the various sensors based on firing cycles. The knock sensor in particular is checked every approx 250 cycles, so a problem should be detected by the KLR within 15-20 seconds after startup.

Also, sending you a PM.

Van

I have an update for those interested...

Installed new knock sensor - did NOT fix the problem.

I do have more, although perhaps not anymore useful, information.

It is doing the full-blown DME driven fuel cut-out, 0.3 bar boost max safety protection - the one that violently shuts off the fuel when you step on the gas. However, there are NO blink codes from the KLR (I tested by unplugging the TPS, and that did successfully throw the 4-1 code).

I checked the mass air flow meter: +5V power - good
Low voltage across potentiometer with barn door shut - within spec
High voltage with barn door fully open - within spec
Resistance of air temp sensor cold - within spec
Resistance of air temp sensor engine hot - ??? haven't been able to test yet

Full load signal from KLR to DME: engine running - +5V (within spec)
engine at full load - ??? haven't been able to test yet

I've wired a little buzzer up to the wires from the KLR to the cycling valve - I'm hoping it'll buzz when I test drive it. That way I can confirm that the KLR is trying to actuate the wastegate.

Van

An update: the little buzzer on the cycling valve worked nicely. It tells me that the cycling valve is doing its thing. But I still have the overboost protection problem.

A friend was nice enough to loan me another KLR - same problem... So it's not my KLR (which was the poll leader!)

I just spent some time with a multimeter checking KLR and DME connections. Everything checks out (grounds and power feeds). The coolant temp sensor also measures good.

I've verified that my dash boost gauge is pretty accurate.

I've put on a little valve that I can use to set the max boost pressure. (When the cycling valve is open – which allows boost pressure to open the wastegate – some boost pressure is diverted back to the J-boot. By controlling HOW much pressure gets diverted, I can effectively control how open the WG is.)

At 1.6 bar absolute manifold pressure (about 9 psi of boost), I can still get the car to go into overboost protection on a long wide-open-throttle pull (from 2500 RPM in 4th gear to redline - at about 5k or 5500 RPM, the DME will start cutting fuel and spark, and retard the timing, even though the boost doesn't go above 1.6 bar).

Maybe it's a bad DME? Anyone have any other ideas?

Van

Solved my problem - after trying another air flow meter, and still encountering the problem, I swapped my DME for a different one. Problem Solved!! Turns out my DME sh|t the bed.

Now the car goes like stink. In 4th gear it gives me full boost all the way to redline. (I didn't have the ***** to try 5th to redline...)

reno808

lol nice

Van

I heard from the place rebuilding the DME - they said, "Problem found in resistor R409".

Luis de Prat

Who's rebuilding your DME, if you don't mind my asking?

Van

I'm using Specialized ECU: http://www.ecudoctors.com/

$350 for the repair, return shipping and a 5 year warrranty... I won't get to test out their work until the end of the month - but I'll be sure to let you know what the outcome is!

Luis de Prat

Thanks for the referral. I hope it works out!